Food packaging production date/batch OCR character recognition
-
Project Description
- High -speed flying identification, read 250 products in one minute;
- The recognition rate is high, reaching more than 99%;
- The surface of the packaging surface is serious, and the optical lighting requirements for the collection of character information images are high.
-
Characteristic
- Customized optical solutions to solve the high -light material surface light reflective phenomenon, which greatly enhances the effect of visual imaging image image;
- Adopt the intelligent deep learning algorithm architecture to automatically extract image features. Only dozens of pictures can reach more than 99%of the recognition rate effect;
- Adopting the imitation transformation algorithm, the automatic correction is skewed, and the self -developed OCR identification software is used to achieve only 200ms from the collection map to the result.